1. Field of Art
The present invention relates to communications networks and, more specifically, to methods and apparatuses for disseminating link state information throughout a communications network and using such information to make dynamic, end-to-end routing decisions for video, voice and data transmissions.
2. Related Art
Many of the conventional routing protocol standards for communications networks require that each node in the network advertise to all other nodes in the network a set of link characteristics, or attributes, associated with each link connected to that node. These link characteristics are commonly known as the “link state” characteristics. Among other things, the link state characteristics may include, for example, routing-related parameters, such as a link metric for a shortest path calculation, a maximum bandwidth, a maximum reservable bandwidth, and/or an amount of unreserved bandwidth. These link state characteristics are typically used by one or more routing algorithms in the system to make dynamic routing decisions (a process known as “traffic engineering”) for each data transmission path across the network.
The typical way of tracking, storing and disseminating link state characteristics in a communications network is to use what is known as a link resource record (sometimes referred to as a resource class identifier, a link color bit mask, or link color bit vector). Currently proposed standards for link resource records typically comprise a 4-byte (or 32-bit) bit-mask-encoded record (or vector) that may be used by a service provider to categorize each link in the provider's network into thirty-two distinct classes. The 32-bit bit-mask-encoded record is supposed to facilitate faster and more efficient provisioning of transmission services, as well as more intelligent routing decisions based on sophisticated service and transport constraints. Since conventional standards do not restrict the actual definitions for the thirty-two classes that can be represented by the bit-mask-encoded record, network administrators may define their own sets of classes.
While a bit-mask-encoded resource record of 32 bits may be adequate for many of today's communications networks, there are some disadvantages associated with using it that are becoming more and more significant. For example, it is widely believed in the telecommunications industry that future multiple services (i.e., data, voice, and video) communications networks (MSCN) will be comprised almost entirely, if not entirely, of fiber-optic transmission equipment. MSCNs that include fiber-optic technology are usually called optical transport networks (OTNs). Conventional OTNs usually are not entirely optical in that they typically include electronic signal processing equipment, such as optoelectronic switches, at nodes in the overall network topology that become intermediate nodes in a given communications path. An optoelectronic switch receives an optical signal, converts it to an electronic signal, performs local switching in the electronic domain, and converts the electronic signal back to an optical signal at an egress port before sending the signal out to the next optoelectronic switch on the path. This process is called OEO conversion. OTNs that utilize electronic signals and OEO conversion during transmission are often called “opaque” OTNs.
Pure OTNs, however, do not convert optical signals to electronic signals during transmission. In a pure OTN, the signals remain entirely in the optical domain from source to destination. Pure optical transport networks are usually referred to as “transparent” OTNs.
The majority of OTNs in operation today are opaque. However, the industry as a whole is shifting away from opaque OTNs and toward transparent OTNs because, in a transparent OTN, optical signals do not need to be processed (e.g., no OEO conversion) as they pass through intermediate nodes. Thus, transparent OTNs typically operate faster, require less equipment than opaque OTNs, and cost much less to build.
To address concerns that their current opaque OTNs are not cost effective and do not benefit from today's rapidly-occurring advances in optical technologies, many carriers in the telecommunications industry are attempting to drive down capital expenses and to “future-proof” their networks by deploying phototonic cross connect (PXC) switches (often described as optical wavelength switches) and hybrid switches in their OTNs. A PXC acts as bridge in a communications path of an optical network, linking light signals of different wavelengths together at the node where two networks meet. A hybrid switch, on the other hand, includes in one box the functionality of both a PXC and an optoelectronic switch. Thus, a hybrid switch can be configured to simultaneously process electronic signals and perform OEO conversion, as well as connect and relay optical signals as would be performed by a regular PXC. The purpose of the optoelectronic functionality in a hybrid switch is to convert locally-originating electronic signals to optical signals for delivery over an OTN, and conversely, to convert optical signals received from an OTN into electronic signals for delivery to one or more electronic destinations.
PXCs are frequently interconnected with dense wavelength division multiplex (DWDM) line systems. In some implementations, DWDM line system functionality is integrated with a PXC in one device called a wavelength crossconnect (WXC). DWDM is a technology that puts data from different sources together on a single optical fiber, with each signal being carried simultaneously on its own distinct optical wavelength. Using DWDM, up to 128 (and theoretically more) separate wavelengths, or channels of data, can be multiplexed into a lightstream, transmitted on a single optical fiber, and de-multiplexed back onto separate optical fibers at the destination end of the transmission path. Since every wavelength (and, hence, every channel) is de-multiplexed onto a separate channel at the destination end, different data formats being transmitted at different data rates may be transmitted over a single optical fiber simultaneously. Thus, Internet Protocol (IP) data, Synchronous Optical Network data (SONET), and asynchronous transfer mode (ATM) data, for example, can all be transmitted within the same optical fiber at the same time.
DWDM, sometimes called wave division multiplexing (WDM), is widely expected to solve the bandwidth exhaustion problem associated with some fiber-optic networks. In a system like Lucent's LambdaXtreme, for example, which supports 128 wavelengths with each wavelength carrying a signal at 10 Gbps, up to 1.28 Tetra Bits of information can be delivered per second by a DWDM-enabled optical fiber. Accordingly, PXC and DWDM promise to be two of the key technologies used in the all-optical networks of the future.
Optical links in a transparent OTN, especially optical links controlled by DWDM technology, have many more link state characteristics associated with them than optical links used in an opaque OTN. Therefore, when the OTN is a transparent OTN incorporating optical links, DWDM and PXCs, there are many more link state characteristics that must be disseminated throughout the network and used by the routing algorithms to make dynamic routing decisions. Moreover, in order to comply with numerous standard routing protocols, each PXC in an OTN must advertise to each other PXC in the network all of the link state characteristics associated with each link connected to the PXC, including, for example, optical fiber types, wavelengths and wavelength bands used on the link, the type of dispersion compensation technique used on the link, a DWDM line encoding scheme associated with the link, and so on.
As stated above, the 32-bit bit-mask-encoded resource record has been proposed for use as an industry standard. Three such proposals are presented, for example, in Katz, D., et al, Traffic Engineering Extensions to OSPF, Internet Draft, draft-katz-yeung-ospf-traffic-09.txt, Li, T. and Smit, H., IS-IS Extensions for Traffic Engineering, Internet Draft, draft-ietf-isis-traffic-04.txt, and Fredette, A. and Lang, J., Link Management Protocol (LMP) for WDM Transmission Systems, Internet Draft, draft-ietf-camp-lmp-wdm-wdm-01.txt, which may be accessed on the Internet Engineering Task Force's (IETF) website at http.//www.ietf.org/ietf/lid-abstracts.txt. All of these references are incorporated herein in their entirety by this reference.
However, the 32-bit bit-mask-encoded resource record restricts the maximum number of classes (or categories) to thirty-two, which are far too few categories to adequately describe the numerous characteristics associated with links in certain types of networks, such as, for example, a typical OTN. In order to address this restriction, some of the hitherto proposed schemes introduce a separate record or object to represent each optical link characteristic. In such a cases, however, each new record or object requires more bits to represent the same amount of information, additional bandwidth to disseminate the additional bits throughout the network, and additional processing power to process the additional bits. Thus, schemes that rely on defining new records and objects for each type of characteristic are inherently less efficient.
In the interest of addressing these and other problems associated with using conventional link resource record systems, what is needed is a more efficient and less demanding method of representing and disseminating potentially hundreds of thousands of link state characteristics that need to be represented in OTNs with DWDM-controlled links. Such a system would be even more useful if it were adapted for use in conjunction with standard link state routing protocols, such as Open Shortest Path First (OSPF) protocol, Constrained Shortest Path First (CSPF) protocol, Intermediate System to Intermediate System (IS-IS) protocol and Private Network-to-Private-Network Interface (PNNI) protocol.